I agree. Can you robustify the analysis by looked at other sample periods? In particular, what happened in a rising rate regime?